- 10 Apr, 2018 1 commit
-
-
Alejandro Sanchez authored
Bug 5007.
-
- 06 Apr, 2018 2 commits
-
-
Tim Wickberg authored
-
Tim Wickberg authored
Reformat comment block while here, and mention TRES as well.
-
- 05 Apr, 2018 1 commit
-
-
Morris Jette authored
Bug 5007
-
- 04 Apr, 2018 1 commit
-
-
Morris Jette authored
-
- 03 Apr, 2018 2 commits
-
-
Ben Matthews authored
Bug 4952.
-
Ole H Nielsen authored
Bug 5011.
-
- 02 Apr, 2018 1 commit
-
-
Alejandro Sanchez authored
If an update for a job that is being requeued[hold] happens, we don't want to modify the start_time of the old record. Pending + Completing is equivalent to Requeue. Bug 4855.
-
- 30 Mar, 2018 1 commit
-
-
Tim Wickberg authored
-
- 27 Mar, 2018 14 commits
-
-
Isaac Hartung authored
sprio truncates the username field to 8 characters, leading to a failure if you run the testsuite with a longer username. Bug 4801.
-
Marshall Garey authored
Remove partition time limits, run the test, then restore partition time limits with reconfigure. Bug 4276.
-
Marshall Garey authored
Remove partition MaxTime limit at the beginning of the test, run the rest of the test, then restore the partition configuration with scontrol reconfigure. Bug 4915.
-
Marshall Garey authored
Bug 4915.
-
Marshall Garey authored
Bug 4915.
-
Marshall Garey authored
Bug 4926.
-
Marshall Garey authored
Bug 3567.
-
Brian Christiansen authored
Missed from 500e23b5 Bug 4988
-
Brian Christiansen authored
job_ptr was still being referenced after releasing the read locks. Also prolog_running_decr() needs to have a job write lock and a fed read lock. Bug 4947
-
Isaac Hartung authored
Bug 4984.
-
Dominik Bartkiewicz authored
Bug 4990.
-
Felip Moll authored
-
Felip Moll authored
When a capmc program is stuck a node can remain in IDLE*+POWER state despite the admin manually rebooted and cleaned it. The status can be changed then by issuing scontrol update ... state=power_down. Added this example in docs. Bug 4788
-
Alejandro Sanchez authored
Bug 4411
-
- 26 Mar, 2018 1 commit
-
-
Felip Moll authored
While restarting slurmctld with -R flag, features were emptied. This patch just does nothing when recover=2 (-R) which means active and available features are preserved from the state files. Needed for backup slurmctld controller. Bug 4927
-
- 23 Mar, 2018 6 commits
-
-
Boris Karasev authored
Bug 4882
-
Felip Moll authored
-
Felip Moll authored
When setting environment variables in jobs we were tracking other job's indexes and information, while the scope of the environment is totally different for each job. The static variables have been moved to non static. This fixes CUDA_VISIBLE_DEVICES indexing while running different jobs on the same node. Bug 4851
-
Tim Wickberg authored
If block_map_size is zero (as seen on FreeBSD when not using HWLOC), xmalloc(0) will return NULL. Writing to str[0] then is invalid. Bug 4845.
-
Tim Wickberg authored
Commit 943c4a13 introduced _register_oom_notifications() which relies on eventfd() to subscribe to the memory.oom_control file. As eventfd is unavailable on FreeBSD, disable this functionality until some alternative implementation can be created. Bug 4845.
-
Tim Wickberg authored
Copy this definition from src/common/gres.c, along with the necessary includes. (Note that these includes cannot be alphabetized as is done elsewhere.) Bug 4845.
-
- 22 Mar, 2018 2 commits
-
-
Brian Christiansen authored
Bug 4971
-
Morris Jette authored
bug 4932
-
- 21 Mar, 2018 3 commits
-
-
Felip Moll authored
-X doesn't show cumulative statistics for each job but statistics relevant to the job allocation itself. Bug 4967
-
Felip Moll authored
It fixes a scontrol segfault when updating a reservation with empty users or accounts and the parameters were specified more than once. Bug 4945
-
Marshall Garey authored
Bug 4778
-
- 20 Mar, 2018 5 commits
-
-
Morris Jette authored
Bug 4949
-
Tim Wickberg authored
-
Tim Wickberg authored
-
Tim Wickberg authored
-
Tim Wickberg authored
-